Looking at the (limited) docs for the demo Fraunhofer encoder (V3.1 Demo (build Sep 23 1998)), one of the options is to enable/disable "Soft time-domain filtering" which they say uses "a high-quality time domain filter instead of fast MDCT". They must be referring to the psycho acoustic model, since it seems the encoding standard would mandate using MDCT for encoding. After looking at a few sharp transitions in some of my own samples, and in castanets.wav (btw, the links on www.sulaco.org are broken), it appears from MP3x that FhG does much better eliminating pre-echo. The audio difference is fairly minor, but noticable. Has anyone experimented with incorporating a time-based filter to see if this would help the pre-echo? Thanks, John -- MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
